Kevin Mitchell wrote an essay this summer (“The genetics of stupidity”) that got some attention. The idea is that most or all of the genetic influence affecting intelligence is genetic load. Although each deleterious mutation is very rare, the average individual carries many of them – and some people carry more than others. People with more messed-up genes than average would be less smart than average.
This has to be part of the truth, and it might be most of the truth, particularly within a population. People have had no success looking for single alleles that explain much of the variation in IQ – but as far as I am aware, all of those studies have been of Europeans. In the next population you look at, the genetic architecture might be different. For example, height is highly polygenic in Europeans. No single allele explains much of the variance, although a few have large-enough influence to be detectable (unlike IQ, thus far). But the situation is different in Pygmies. There, we see loci that influence height, as well as a general tendency of height increasing with the fraction of the Bantu ancestry.
So you don’t know until you look. It could well be that there are IQ QTLs among the Pygmies: if nothing else, being bigger, with a bigger brain, might well increase IQ. If I had to guess, I would suggest that IQ in Pygmies might also increase with the fraction of Bantu ancestry, which, if true, would certainly mystify many people.
I doubt if genetic load is the entire story. Selection happens. But load surely matters, and may be part of the explanation both within and between populations.
Years ago, I don’t think it was obvious that you wouldn’t find a few fair-sized IQ QTLs in Europeans. It was obvious that genetic load would be part of the story – I’ve been thinking about it for a long time. There I go again, using that mysterious and undefinable word again – ‘obvious’. Have to be careful about that.
West Hunter 
Do you have any insights into why there is such a large variation in actual human intelligence when the genetic evidence to date can’t find anything that would account for this. Using height for an example it is as if intellectually people vary in height from four feet to ten feet yet human intelligence is far more polygenic than height. I just don’t get it, with all these tiny tiny genetic variables influencing intelligence we shouldn’t have the variation that exists.
I bet you’re thinking it’s like a huge pile of coin flips => sharp gaussian. But with more genes, the range of possibilities gets larger, and if there’s any push / selection there are more places to go.
I think Mitchell is wrong, that genetic load does not account for the majority of variance in human intelligence. If he were correct than the standard deviation in human intelligence would be very small, but in fact it is very large. What causes it? Who knows, I hope to live long enough to see companies like BGI shed light as to why.
It depends on how IQ varies with the amount of load. The semi-educated guess is that the effects of load are multiplicative, which means that the observed differences in IQ could be explained by differences in load.
A part of your statement sounds very likely to me. That are yet to be discovered combinations of genes or the switching in front of the genes that are multiplicative and this accounts for the wide range in human intelligence. Genetic load combinations are very likely to account for people on the wrong end of the bell shaped curve. But what about the other end. Isn’t it more likely that their high intelligence is the result of the right combination of multiplicative genes rather than the lack of the wrong combination.
You’re talking about epistasis, but I was not. I was talking about how a trait might change as the amount of genetic load increased. The usual model has fitness (a special trait) declining exponentially with the amount of load.
I’ll repeat the question I asked in a different threat which received no response (I’m not bitter):
So, if differences between groups are due to genetic load, will this make it more difficult to show scientifically that there are genetic differences in IQ between groups?
I had hoped that science would detect genes that influence IQ whose frequencies differed between groups so that the debate could finally move forward. What are the odds of the debate moving forward under the genetic load as primary determinant scenario?
Genetic load probably won’t be too hard to figure out. If it turns out to explain a significant fraction of intergroup differences, it won’t ‘move the debate forward’ one iota.
Why not?
Why not? Because the inter-group differences I presume you’re talking about are already obvious, ubiquitous, and persist from generation to generation. Yet no one admits they even exist, never mind have any sort of hereditary basis. How will some arcane evidence regarding genetic load convince anyone otherwise?
Right. I figure that clear genetic evidence would change the minds of hundreds of people worldwide. Hmmm. So what is the point of knowing? Well, what I said is not forever and universally true. Occasionally the world goes through some sort of phase change, and it turns out that no one really believed what everyone said and did a week ago. Or so they say now. Such reversals can be mystifying, even hilarious, when you happen not to be part of them. Sometimes they have their roots in something practical. For example, if understanding of the genetic roots of intergroup differences led to an effective intervention, attitudes might change.
Perhaps when Steve Hsu throws out his current BGI data I will now be qualified to contribute to his new sample.
Pingback: linkfest – 11/12/12 « hbd* chick
http://medicalxpress.com/news/2012-11-humans-slowly-surely-intellectual-emotional.html – Human intelligence and behavior require optimal functioning of a large number of genes, which requires enormous evolutionary pressures to maintain. A provocative hypothesis published in a recent set of Science and Society pieces published in the Cell Press journal Trends in Genetics suggests that we are losing our intellectual and emotional capabilities because the intricate web of genes endowing us with our brain power is particularly susceptible to mutations and that these mutations are not being selected against in our modern society.
Provocative? I won’t speculate on time frame, but it’s obviously going to happen if you’re not constantly undergoing selection in the opposite direction.
This is late in the chain, so forgive me if the original issue has now been forgotten. Dave Chamberlain asked “Do you have any insights into why there is such a large variation in actual human intelligence when the genetic evidence to date can’t find anything that would account for this”. I agree that the variation is large, but compared to what? Height might not be the best comparison. Perhaps muscle strength or running speed might be better (more easy to vary this than the whole skeletal structure). Art Jensen said that in a standard classroom there would always be one child who could complete a task 5 times faster than the slowest child. That is a big difference, but Ian Deary et al are on the way to “explaining” about 45% of that variance, if only in the sample of discovery
I found it difficult to understand how LoF mutations might have an impact on IQ based on the following simple model.
Assume that 150 genes are involved in IQ, all of which have two alleles, where one allele of the two contibutes 1 IQ point and the other contributes 0 IQ points. In this model, IQ is a binomial distribution and ranges from 0 to 150 with an average of 75. Pretty bad model because an IQ of zero seems impossible. So, we could assume that 50 of the genes are fixed at the 1 IQ point allele, and then IQ would range from 50 to 150 with an average of 100. We can play around with other sets of genes fixed on one allele or the other and can get a feel for some interesting things, although I am sure it is very simplistic.
Anyway, I couldn’t figure out how LoF might affect IQ until this morning in the shower when it struck me that FI LoF alleles will affect IQ, where FI is the fraction of genes involved in IQ, and in my simple model if they affect genes that contribute 1 IQ point, they they will reduce the IQ of the individual who inherits them, but if they affect genes that contribute 0 IQ points, well, as I said, pretty simple model.
Have I got the wrong end of the stick?
Pingback: Features and Bugs | JayMan's Blog
Pingback: Jayman’s HBD Reading List | Propertarianism